This essay is an outgrowth of an effort to determine the phylogeny of the families of the Lepidoptera, in order to decide upon a classification to be used in a general text book of Entomology. More than three years were devoted to the problem before a systematic method of procedure was adopted. This time was largely spent in a comparative study of published classifications and in an effort to determine which of these represented most accurately the facts of nature. A large part of the work yielded poor returns for the labor expended; for it was carried on with no definite plan; it was a blind groping in the dark.

Suddenly one day a flood of light was thrown upon the work by the recognition of the fact that a moth which I was studying (Hepialis, Plate I, Fig. 2) was a generalized type. I found that a knowledge of the structure of this insect gave a clue to the probable structure of the primitive Lepidoptera. And that with this knowledge it was not too much to expect to be able to trace out the various lines of descent represented by existing forms. Then began a systematic study which has resulted in the development of the method outlined in Part I of this essay.

I regret that I have been unable to apply the method as fully as I should like to before publishing it. But the results which I have been able to obtain by it lead me to hope that the publication is not premature. And as the leisure which a teacher can get for study is limited, I could not hope to make a complete application of it, even to the families of a single order, for many years.

A complete application of the method to the Lepidoptera alone will involve a study of the segments of the body as a whole, the peculiarities of development of particular segments, the structure of internal organs, the structure of organs of special sense, the various appendages of the body as antennae, mouth-parts, legs, wings, and the external appendages of the reproductive organs, the clothing of the body, in a word the study of every structure that has been developed in these insects.

But although this extended study will be necessary before we can consider our work complete, a provisional classification can be based on the study of a single organ or set of organs. We have only to remember that such a classification is merely provisional, and that the results obtained in this way should be confirmed or corrected by the study of other organs.

Following the method indicated, the wings were selected as the first organs to be studied. These organs were chosen as the most available ones, owing to their size, and the ease with which variations in their structure could be observed. The record of the action of natural selection is recorded upon them as upon a broad page. For a long time my attention was confined to the wings of the Lepidoptera. But later I found it necessary to greatly extend my studies in order to determine the primitive type of the wings of insects. It was also necessary to study the wings of insects of other groups in order to select a nomenclature of the wing veins that would apply to all orders of insects.

Although there are great differences in the venation of the wings of insects of different orders, a study of the more generalized members of the several orders of winged insects show that the type of venation is the same for them all. This indicates two points of great scientific interest: first, wings have originated but once in the class Hexapoda, or, to state the same thing in other words, all of the orders of winged insects have descended from a common stock; second, if all the various forms of wings are modifications of the same type, it is not too much to expect to be able to establish a uniform nomenclature for the principal elements in the frame work of the wings, i.e., the principal veins, although doubtless it will be necessary to use special names in many cases for structures that have been developed secondarily.

The importance of the wings of insects for taxonomic purposes was early recognized by entomologists, as is well shown by the fact that the names of the Linnean orders are all drawn from the nature of the wings, except one, Aptera, and that from the absence of wings.

Although, doubtless, the great extent to which the wings are still used in taxonomy is partially due to the ease with which wing characters can be observed, still the following considerations show that such use is warranted by the facts of nature.

The chief end of existence of an adult insect is to provide for the perpetuation of the species. This resolves itself in the case of the male into seeking a mate; and in the case of the female after accepting a mate, into seeking a proper nidus for her eggs. In the case of certain insects special conditions necessitate a prolonged existence in the adult state in order to accomplish this end in the best manner; in such cases there may exist a necessity for seeking food; but in many families all nourishment is taken during the adolescent stages.

The necessity for seeking mates or for properly placing eggs, as well as for seeking food gives great importance to organs which increase the power of locomotion. It follows from this that when organs of flight had once been developed such organs would furnish an important field for the action of natural selection.

It has been indicated that there is good reason to believe that all winged insects have descended from a common winged ancestor. But we find that the primitive type of wing has been modified in many widely different ways. Hence a study of the various ways in which wings have been specialized can not fail to throw much light on the phylogeny of insects.

The fact that in some cases, notably those of most animal parasites, wings, becoming unnecessary and perhaps even detrimental, have been lost does not lessen the value of these organs for taxonomic purposes when they have been preserved.

It is often urged, that as the wings are merely appendages of the body, they are extremely liable to be modified in form; and that consequently we cannot hope to find in them a very permanent record. In other words, while it is generally admitted that variations in the framework of the wings may give us important clews as to the limits of the genera, we can not hope to base conclusions upon them as to the limits of larger groups. But it does not follow, that because an organ is a superficial one, it is of little value in suggesting broad generalizations. We find that often the most superficial of structures are among those that were developed very early in the history of a large group, and have persisted almost unchanged in form, although more central structures have been greatly and variously modified. Thus the form of mere dermal appendages may present characters of very high value, as the hair of mammals and the feathers of birds; even the Cetacea have hair, and Archopteryx had feathers. As "the proof of a pudding is in the eating," so the value of a character for taxonomic purposes can be determined only by its use.

As to the origin of wings we have no knowledge. We have not even a generally accepted theory to account for the appearance of these structures. Many writers believe that they are modified tracheal gills. We find in many aquatic nymphs plate-like gills, some of which would need to be modified but little to function as organs of locomotion. This is especially true of the covering pieces forming part of the respiratory appendages of an Ephemerid nymph. These pieces not only protect the gills beneath them; but probably also serve by their flapping to cause a current of water to pass over the gills. Fig. 1 represents the covering piece of a tracheal gill of an Ephemerid nymph collected at Ithaca. The step from such a structure to one that would aid in locomotion is not a great one.

But other writers think that the wings arose as keel-like expansions of the sides of the thorax. Such expansions would function as a parachute in a falling insect, as does the folds of skin in a flying squirrel, or would function both as a kite and a parachute in a leaping insect.

In support of the latter theory the netted-veined triangular prolongations at the sides of the prothorax of certain fossil insects (Choraedodis and Lithomantis, Fig. 2) are brought forward; and it is also said that a species of Tingis from Texas shows on the prothorax, transparent projections of triangular form and a netted venation similar to that of the fore wings¹.

Whatever the origin of wings may have been, I think that this much is clear: they were developed to comparatively large size and were furnished with numerous veins before they began to function as active organs of flight. This large size and rich venation may have been the result of a natural selection of those forms best fitted to act as a kite or a parachute. But this type of wing is not well adapted for active flight. As soon as there arose a tendency for the wing to function in this way, there began, doubtless, the extensive series of modifications of which we have records both in the rocks and among living insects.

We know almost as little regarding the origin of the veins of the wings as we do of the wings themselves. Still we may be allowed to speculate regarding the matter. Let us suppose that the wings originated from broadly expanded organs; such organs would be furnished with tracheae, even if they were used only as a parachute; for they would need to be supplied with air as are other parts of the body. On the modification of such organs into wings fitted for active flight, it would be important that the main trunks of the tracheae should be protected in such a way that any bending of the wing, which would compress them and thus choke off the supply of air, would be avoided. Thus any tendency of the membrane of the wing to become thickened along the lines of the tracheae would be preserved by natural selection; and correlated with the development of this firm frame-work, there might be a thinning of the spaces between the main tracheae, thus insuring lightness of the entire organ.

We are not entirely without evidence that this is the method of the formation of wing veins. It will be shown later that when the principal stem of one system of veins (media) is obliterated and the branches of this system are forced to derive their supply of air through lateral tracheae extending to adjacent systems of veins, these lateral tracheae become enveloped by veins resembling in every respect the principal veins. Similar evidence may be drawn from the study of the development of the humeral veins in the Lasiocampidae.

Since we are not able to determine the form and function of the organs which were modified into wings, let us endeavor to select the most generalized type of wings preserved to us. We will first see what light Palaeontology throws upon this question

Comparatively little is known regarding the primitive insects. But thanks to the labors of Mr. S. H. Scudder, who has been one of the chief workers in this field, what is known has been made easily accessible.²

Winged insects appeared very early, probably as early as any land plants; for Moberg has figured an insect (in the Forhandlingar of the Swedish geological society) from the upper part of the lower silurian; and Brongniart has figured and described a wing from the middle silurian sandstone of Calvados, France. (Fig. 3). But this wing instead of being primitive in form represents a rather highly specialized type, if the figure given correctly represents its structure. And the insect described by Moberg (Protocimex siluricus) is supposed by him to be hemipterous, an even more highly specialized type.

Of devonian insects we know several. Those which are best preserved are Homothetus fossilis (Fig. 4), Xenoneura antiquorum (Fig. 5), and Platephemera antiqua (Fig. 6). These differ among themselves to such an extent that we are forced to conclude without taking into account the two known silurian insects, that already at that early time there was a large and varied insect fauna, of which the more primitive forms have not been discovered.

From the carboniferous rocks much more abundant material has been obtained. But, according to the views of Mr. Scudder "there existed among these ancient forms no ordinal distinctions, such as obtain today, but they formed a single homogeneous group of generalized hexapods, which should be separated from later types more by the lack of those special characteristics which are the property of existing Orders than by any definite peculiarity of its own." ³

To this group of generalized hexapods which includes all paleozoic insects the name Palaeodictyoptera has been applied.

Among the Palaeodictyoptera were insects which were undoubtedly the precursors of the cockroaches, the may-flies, and the walking-sticks. Still these groups of insects "were more closely related to one another, at least in the structure of their wings (which is the only point of general structure yet open for comparison) than any one of them is to that modern group to which it is most allied."  The ordinal distinctions which is now found in the "wing structure of modern insects did not exist in paleozoic insects, but a common simple type of venation which barely admitted of family divisions." ⁴

In his classification of the Palaeodictyoptera,⁵ Mr. Scudder indicates a number of these family divisions, and groups them according to their general facies and by their relationship to succeeding types into four sections as shown bv the following table.⁶
 

As this is a comparatively late period we are forced to pursue practically the same method that we would were we to attempt to solve the problem by a study of living insects. That is we must select what seems to be the more generalized types and try to eliminate from these the results of sidewise developments.

What is gained by the study of carboniferous insects, I believe, is the demonstration that certain characters which have been commonly considered primitive did not exist with these insects, and consequently must be considered as secondary developments.

A comparative study of these insects show that in the Palaeodictyoptera the two pairs of wings were very similar in structure both being membranous and furnished with a similar framework of veins.⁷ Each wing possessed six principal veins or groups of veins which are clearly homologous with the costa, subcosta, radius, media, cubitus, and anal veins of modern insects. The wings when not in use were closed over the abdomen. They were sometimes broadly folded but were never plated, as are the hind wings of the Acri-didae for example. This feature being with little doubt a comparatively late development. (See Scudder, Pretertiary Insects, p. 49.)

It seems to me probable that the Palephemeridae presented an exception to the general statement given above, in that the two pairs of wings were not of equal size. The shape of the fore wings of Platephemera antiqua (Fig. 6) is such as to indicate that the hind wings were small. And I am informed by Mr. Scudder that the only paleozoic may-fly known that shows the hind wings at all (Palingenia feistmanteli) shows that they must have been broad and in all probality much shorter than the fore wings.⁸

Although the wings of the Palaeodictyoptera agree in the characteristics given above they present great differences of structure. How shall we decide which of the different forms is the more primitive. This I think can be done by determining the degree of their adaptation to the performance of their function.

In those insects which have the best developed powers of flight we find that the costal edge of the front wings is the strongest part of the wing, the strength being due to the presence of several strong veins which are nearly parallel. Such all arrangement is necessary to withstand the strain that is brought upon this part of the wing. I conclude, therefore, that wings possessing this structure are more highly specialized than those in which the costal edge is not strengthened in this way. Using this criterion I select that form of wing which departs most widely from this type as the most generalized form; for so important a character as this when once attained would not be lost so long as the wings were used as organs of vigorous flight. I conclude, therefore, that it is among the carboniferous cockroaches that we find the most generalized form of wing.

Owing to our limited knowledge of extinct forms, I cannot hope to present in these studies continuous series; but can only select examples which illustrate the direction or directions of specialization of particular parts, without strictly confining myself to a single line of development.

In Mylacris anthracophilum (Fig. 7) the wing is furnished with five sets of veins in addition to the costal or marginal vein if this exists. Three of these which occupy the intermediate area of the wing arise each in a single strong trunk. These veins may be designated beginning with the one nearest the costal margin as the radius, media, and cubitus, respectively.

Lying between the costal edge of the wing and the radius there is a group of veins which radiate from near the base of the wing (Fig. 7, II). These I believe represent the subcosta in its most generalized form. These subcostal veins resemble very closely in form and arrangement the group of veins lying behind the cubitus, the anal veins. In fact a longitudinal line drawn through the center of the wing divides it into two nearly similar halves.

It will be readily seen that this type of wing is poorly fitted for active flight; the costal edge lacking the strength necessary for this purpose. In fact the arrangement of the veins approximates that of the covering piece of a tracheal gill figured above (Fig. 1), or that of the lateral appendages of the prothorax of Lithomantis (Fig. 2).

In Necymylacris heros (Fig. 8, II) we see the beginning of a strengthening of the subcostal area. One of the subcostal veins, the hindermost, becomes the principal vein of this area; and most of the other subcostal veins have become consolidated with this one, so as to appear to be branches of it. The subcostal area is also relatively much narrower.

In Ectoblattina lesquereuxii (Fig. 9) the tendency of the many sub-costal veins to become consolidated into a single strong vein with branches extending to the costa is carried much farther; and the narrowing of the subcostal area is also more marked.

The step from the form of the subcosta in Ectoblattina to that presented by many modern insects is not a great one, as for example, that of the fore wing of Corydalis.

It is probable that correlated with the lengthening and narrowing of the subcostal area in these paleozoic insects a thickening of the costal edge of the wing took place, thus forming the costal vein. As I have been able to study only figures of these paleozoic remains, I am unable to decide at what point in the development of the wing a distinct costal vein was formed. It was probably very early; for Mr. Scudder states that in the paleozoic insects the six principal veins were always developed, the marginal [costa] simple and forming the costal border.

From this brief study of the development of the subcostal area let us pass to the area lying next to the opposite margin of wing, the anal area, omitting for a time any discussion of the three veins (radius, media and cubitus) which occupy the central portion of the wing.

A striking feature in the structure of the wings of many insects is the separation of the anal area from the remainder of the wing by a fold or furrow, along the bottom of which extends a vein. Such a depressed vein has been termed, on account of its position, a concave vein; and in contradistinction to such veins, those veins which extend along the summit of ridges, or which are more prominent on the upper surface of the wing than on the lower, are termed convex veins.

This furrow separating the anal area from the preanal portion of the wing appeared very early. It is especially prominent in all cockroaches both fossil and living; and can usually be recognized in any insect wing in which the anal area is well developed. I have been unable to determine the significance of it. But have found it a very useful mark in defining the limits of the anal area. It is vein VIII of the nomenclature adopted in this paper.

The primitive form of the anal area is probably well shown in Mylacris, (Fig. 7), where it closely resembles the primitive form of the subcostal area, as shown in the same genus. But the latter specialization of this area has been very different from that of the subcostal. This specialization has taken place in two opposite directions, i.e., by reduction and by addition.

In certain lines of development the tendency of natural selection has been to preserve the narrower winged forms. And the narrowing of the wings has taken place largely through a partial or complete reduction of the anal area. The dragon-flies, Odonata, and the ant-lions, Myrmeleon, are examples of the extreme result of this tendency. And in the Lepidoptera there are several instances where a good series illustrating successive stages in this reduction can be found. Thus within a single family, or perhaps superfamily, the more generalized members have three anal veins in at least one pair of wings, (usually the hind wings), while as one passes to more and more specialized forms only two, or one anal veins are found.

I believe that this selection of the narrower winged forms is the result of the survival of those forms that are best fitted for rapid flight. A good illustration of the difference in the powers of flight between an insect with a wide anal area and one in which this area has been reduced, can be found within the limits of a single lepidopterous superfamily, the Zygaenina. Compare the power of flight of Triprocris (Fig. 10) in which there are two anal veins in the fore wings and three anal veins in the hind wings, with that of Syntomis (Fig. 11) in which there is only a single anal vein in both fore and hind wings.

On the other hand, in other lines of development, natural selection has evidently tended to a preservation of the wider winged forms; and the widening of the wings has taken place largely by a specialization of the anal area by addition. The extreme result of this method of specialization is presented by the Orthoptera and especially by the hind wings of the Acrididae. Here we find a widely expanded anal area, with regularly alternating concave and convex veins. Such a wing is not fitted for striking vigorous and rapid blows upon the air as is required for rapid flight; but is adapted to a sliding flight, a sliding up like a kite or down like a parachute. Such a method of flight would naturally reach its highest development in jumping insects, like the Acrididae.

A study of the illustrations just given shows that where the tendency of natural selection is towards the development of a rapid flight there is usually a cephalization of the function of flight, i.e., the hind wings are greatly reduced, and the fore wings become the chief organs of flight. This is well shown by the more specialized Zygaenids (Fig. 11); and the extreme of such a cephalization is presented by the Diptera. That such a cephalization is not absolutely necessary to rapid flight is shown by the dragon-flies (Odonata); but here the abdomen is greatly elongated, which gives a similar result.

On the other hand where an expansion of an anal area has taken place in order to provide for a sliding flight, it is the hind wings that are specialized by addition, i.e., the opposite of cephalization takes place. The Acrididae have already been cited as an illustration of this.

The region lying between the subcostal and anal areas is traversed by three principal veins and their branches. These veins as already indicated, are the radius, media and cubitus, the radius lying next to the subcosta, the cubitus next to the anal area, and the media, between the radius and cubitus. Very remarkable modifications take place in the structure of these veins and in their relation to each other. Some of the modifications will be discussed in detail later; in this place I wish only to make some very general statements.

If a large series of wings be examined it will be found that the area of each of these veins may be specialized either by addition or by reduction, i.e., it may be either widened or narrowed. When the tendency of natural selection is to widen one of these areas, the points of origin of the branches of the principal vein will be nearer the base of the wing in the more specialized forms than in the more generalized members of the same group, On the other hand when the tendency of natural selection is to narrow one of these areas the branches become consolidated with the main stem to a greater and greater distance from the base in the more and more specialized forms. This consolidation of a branch with the main stem or of two branches with each other may extend to the margin of the wing, and thus the number of branches be reduced. This migration of the point of origin of a branch of a vein often affords an excellent clew to the degree of departure from a more generalized type.

But the most remarkable of the changes which take place in this region of the wing is an abortion of the main trunk of media and a consequent uniting of the branches of this vein either with cubitus or with both cubitus and radius. Excellent illustrations of this occur in the Lepidoptera. In many of the more generalized moths the main trunk of media is well preserved (Fig. 12): while in more specialized forms it is entirely wanting. Sometimes, as in Danais, remnants of the basal part of the branches of media project back into the disca1 cell from the distal vein (Fig. 13); while in many other butterflies the branches of media are so completely united with radius and cubitus that there is no indication of the fact that they do not belong to these systems of veins (Fig. 14).

It is probable that in none of the Pa1aeodictyoptera were the wings plaited, as they are in many existing insects; although in some, they were broadly folded. And if we except the anal furrow (vein VIII), already referred to, all of the veins were of the type that is termed convex; that is, they were more prominent on the upper surface of the wing than on the lower.

We thus see that the evidence of the Palaeodictyoptera does not corroborate the theory of Adolph and Redtenbacher as to the primitive type of the wings of insects. Instead of the primitive wing consisting of regularly alternating concave and convex veins, as described by there, it is probable that the concave veins are a later development, either arising de novo or being modified convex veins, excepting always the anal furrow (vein VIII), regarding the origin of which we know nothing.

Concave veins have evidently arisen to meet two distinct needs: first, in those insects in which the wings have become broadly expanded so as to provide for a sliding flight, there is a necessity for the plaiting of these wings when not in use so that they may be carried without impeding locomotion on foot; second, we find in certain cases where the tendency of specialization has been towards a narrowing of the wings in order to admit of vigorous flight, a corrugation of the wings has taken place in order to strengthen them. The hind wings of a grasshopper illustrate the first; and the wings of a dragon fly present the extreme of the second form of specialization.

It is easy to see that a corrugated wing, like that of the dragon fly, is much stiffer than it would be if the membrane extended in a single plane. If one will examine the cross veins extending between the costa and the radius in a dragon fly, he will find that some of these are in the form of triangular braces which effectually prevent any tendency on the part of the wing to become flattened. Evidently the corrugation is of extreme importance.

The concave veins have arisen in two ways. The first of these is by a change in the position of a convex vein. The subcosta in most of the orders of insects is an illustration of this. In the Lepidoptera the subcosta has retained its form as a convex vein, but in most orders of insects the area between the costa and the radius has been depressed forming a furrow along the bottom of which the subcosta extends. This corrugation has resulted from the need of a stiffening of the costal edge of the wing. The second method of formation of concave veins is illustrated by a vein that lies between radius and media (vein IV), and also by a vein that lies between media and cubitus (vein VI) in certain orders of insects. These veins (IV and VI) I do not believe existed in the Palaeodictyoptera; at least, I have not been able to find any indication of their presence in the figures of paleozoic insects. In the more modern orders of insects when a corrugation of the wings arose, and the areas traversed by these veins became depressed, veins IV and VI appeared. It is probable that they were developed by a straightening out of the zigzag line between two series of cells. This can be readily seen by comparing the wing of one of the devonian may flies (Plate III, 3) with that of a modern may-fly (Plate III, 5). In the devonian may-fly the cells of the wing are polygonal, while in the modern may-fly they are quadrangular. In the latter case not only have longitudinal concave veins been formed from zigzag lines, but the cross veins extending between these concave veins and the adjacent convex veins have become strictly transverse. An arrangement which insures the preservation of the corrugations.

In a similar way the concave veins in the anal area of the modern Orthoptera have probably arisen.

I conclude, therefore, that in the more highly specialized wings of certain orders of insects, there exists a regular alternation of convex and concave veins, this alternation being the result of a corrugation of the wings for the purpose of stiffen-ing them. This conclusion is quite different than that reached by Redtenbacher, who starts with the fan type of wing as the primitive one.

In the Lepidoptera this corrugation has not taken place, the wings being stiffened by scales, consequently, the subcosta remains a convex vein, and veins IV and VI have not been developed. It is probable that these veins are also lacking in the wings of the Hymenoptera and the Coleoptera, but I have not studied carefully the wings of these insects.

As to the nomenclature of the wing veins of insects, there is no longer any doubt regarding the desirability of a uniform system of naming the veins in the different orders of winged insects. Only by such a system can those comparisons be made which are necessary in any thorough study of the relationship of the orders to each other. Heretofore the students of each order have had their peculiar nomenclature, and in many cases writers treating of a single family have proposed a set of names to be used in that family alone. The matter has been further complicated by the fact that not only have different names been applied to the same vein, but the same name has been applied to different veins. Thus the terms costa and subcosta have been applied by Lepidopterists to different veins than those that bear these names in other orders.

There have been several attempts to establish a uniform nomenclature. Of these that of Redtenbacher is the most important, being based on a much more extended study of the subject than that made by any other author.

Redtenbacher was the first one to work out a system and apply it to all of the orders of winged insects. And although his system was based on what I believe to be a false theory, and his interpretation of facts in some cases were faulty, I believe that the more essential features of his system can be adopted.

Although, as I have pointed out above, the fan-type of wing was not the primitive type, it seems desirable to base our nomenclature on this type; for here we find the maximum number of veins; and our nomenclature should include the secondarily developed veins of modern insects as well as the primitive veins.

I have shown that in the preanal portion of the wing of paleozoic insects there were developed five principal veins. These may be designated, beginning with the one on the costaI margin of the wings, as costa, subcosta, radius, media and cubitus. The term media was proposed by Redtenbacher; the others were adopted by him as those sanctioned by the best usage.⁹  I have also shown that in certain insects there  is developed, secondarily, a longitudinal vein between radius and media, for this I propose the term premedia; and also in the same insects there is developed a longitudinal vein between media and cubitus, this I designate as postmedia.

Following the system of Redtenbacher these veins may also be designated by Roman numerals. The equivalence of the numbers and names of the veins of the wing is indicated by the following table; also the nature of the veins.
 

et al.

It will be seen from this table that if we consider subcosta a concave vein, which it has come to be in the larger number of the orders of insects, there is a regular alternation of convex and concave veins, when the maximum number of veins is present. And hence the convex veins are designated by odd numerals and the concave veins by even numerals; this is one of the chief features of Redtenbacher's system, and an exceedingly useful one.

If this system be applied to the anal area, and I believe it is best that it should be, all convex anal veins must be designated by odd numerals. In those orders where the anal area has been greatly specialized by addition, (e. g., Orthoptera), this would naturally follow; for there we find a regular alternation of concave and convex veins. But in certain other cases it is not so obvious. In those Lepidoptera in which three anal veins are preserved, the first (vein VIII) is concave, and doubtless represents the primitive anal furrow; the second is convex and is obviously vein IX; but the third is also convex! Shall this be designated as vein X, or as vein XI? It seems to me better that we consider vein X absent, as are veins IV and VI in this order, and designate this one as vein XI. This view is strengthened by the fact that in many of the Microlepidoptera with broad hind wings there is a prominent fold between the two convex anal veins. This fold may be looked upon as the beginning of an anal vein, which is as yet undeveloped, but which if developed would be vein X.

Another important feature of Redtenbacher's system is the designating of the branches of a vein by Arabic indices appended to the Roman numeral indicating this vein. Thus the branches or radius are designated as III1, III2, III3, etc., (Fig. 15).

While I believe that we are able to trace out homologies between the principal veins of the wings of insects of the different orders, I do not think it practicable, even if possible, which I doubt, to homologize the branches of the principal veins beyond the limits of a single order. I have not, therefore, adopted in all cases Redtenbacher's plan of using odd indices only for convex branches and even indices only for concave branches. This plan will be found very useful in those orders (e.g., Ephemerida) where the fan-type of wing has been developed; but in other cases (e.g., Lepidoptera) it would merely complicate the nomenclature without adding to its value. Thus in the Lepidoptera I designate the five branches of radius as III1, III2, III3, III4, and III5 respectively, although all of these branches are convex.

It should be noted that in numbering the branches of a principal vein, they are numbered in the order in which they reach the margin of the wing, not in the order in which they are given off from the main stem. The system adopted is not only the simpler but insures the same number being applied to homologous veins in different genera, which would not be the case were the other system adopted.

Having indicated the more general features in the development of the wings of insects in order to define the nomenclature of the wing veins that I have adopted, and having explained this nomenclature, I can now pass to the consideration of certain details exhibited by the wings of the Lepidoptera.

The more important of these are the changes which take place in media; for this vein in the Lepidoptera is of the highest value for taxonomic purposes.

The tendency to abortion of the main trunk of media has already been pointed out. The explanation of this tendency I have not fully determined satisfactorily to myself, I can only suggest the following: In the course of the narrowing of the wing and the strengthening of the main veins which has taken place as a result of a natural selection of the more active flying forms, the veins have become crowded together at the base of the wing. The more important veins, i.e., radius and cubitus, have held their place, while media has been crowded out. This crowding out has probably taken place in this way. The narrowing of the space occupied by media compressed the large trachea or tracheae which it contained. Such a compression tends to shut off the supply of air to that part of the wing supplied by the branches of media. To counteract this evil, communication is established between the branch of media and the veins lying on either side of it. When such a communication is well established there is no longer any need for the basal portion of media and it becomes atrophied. In this connection it should be stated that the membrane of the wing is supplied with an immense number of minute traeheae extending from the main trunks contained in the veins. The lateral branches of the traeheae are rarely seen even in carefully bleached wings, for in mounting the specimen they become filled with the mounting medium and are thus rendered invisible. But occasionally air will remain in them rendering them distinctly visibly. It is by means of some of these lateral tracheae that the branches of media become connected with radius and with cubitus.

When such a communication has been established it is important that these tracheae should not be compressed by the bending of the wing during flight, therefore any tendency to protect these tracheae by a thickening of the membrane along their course would be beneficial and would result in the development of veins enclosing these tracheae.

These veins at first extend in a transverse direction, and are thus obviously cross veins (Fig. 16, c.v.). But the result of farther specialization is to round off the angles in the path of the tracheae, as the angles in our roads are rounded off by carts. This process is continued until these cross veins become parts of longitudinal veins, and their true nature as cross veins is completely hidden. This is well shown by the connection existing between the third branch of media (vein V3) and cubitus. A study of the venation of Castnia (Fig. 15) shows conclusively that media is three-branched and cubitus only two-branched. Here the connection between vein V3 and vein VII is obviously a cross vein. But in every American moth and butterfly known to me, except perhaps Hepialis and Micropteryx, the union of these two veins is so complete that there is no hint of the fact that vein V3 is not a branch of vein VII. And in several families vein V2 has also become united with vein VII in a similar manner. The result is that cubitus (the median vein of many authors) is described as three-branched in some families and four-branched in others.

Two years after I had reached the conclusion that media is three-branched and cubitus only two-branched in the Lepidoptera, Spuler published a paper¹⁰ in which these facts are demonstrated in an entirely different way. As I did not publish my conclusions, the credit of the discovery belongs of course to Spuler. I wish merely to state that my conclusions were reached independently of that author's work, and by an entirely different method. I was led to the correct understanding of the relation of these veins by a study of existing generalized forms (especially Hepialis and Castnia); while Spuler's conclusions were based on a study of the ontogeny of certain butterflies. He found that in newly formed pupae the trachea which later becomes enclosed by media is three-branched, while that one which is the precursor of cubitus, is only two-branched. This is an interesting instance of the evidence of ontogeny confirming results obtained in an effort to determine the phylogeny of a group by the study of generalized forms.

¹⁰ A Spuler.--Zur Phylogenie und Ontogenie des Flugelgeaders der Schmetterlinge. Zeit. fur wissenschaftliche Zoologie, LIII. 

Let us see how the facts regarding the changes of media can be used in taxonomic work.

First, the presence of the main trunk of media is an indication of a generalized condition. This at once throws light on the position of the Megalopygidae, the Psychidae, the Cossidae, the Limacodidae, and certain of the Zygaenina. These families are evidently much nearer the stem form of the Lepidop-tera than are those families in which media has been lost.

It does not follow that these families should be classed together. For each one may represent a distinct line of development. The presence or the absence of the base of media is a character that merely indicates the degree of divergence from a primitive type (see p. 43). The divergence in each case may be along a distinct line. It may be worth while to state in this connection that the families named above are nearly all of those of the Macrofrenatae in which three anal veins are preserved in the hind wings, another character indicating a comparatively slight degree of divergence from the primitive type.

Correlated with the abortion of the base of media is the coalescence of its branches with the adjacent veins. It follows from this that the extent to which this coalescence has gone is an indication of the degree of departure of a form from the primitive type. Compare, for example, the hind wings of Packardia (Fig. 16) with the hind wings of Adoneta (Fig. 17), two genera of the family Limacodidae. In Packardis, where a remnant of the base of media still persists, vein V1 is merely connected with vein III by a cross vein. But in Adoneta, where the base of media of the hind wings is lost, vein V1 has become consolidated with vein III for a considerable distance. It is obvious that in these respects, the loss of the base of media and the extent of the coalescence of veins III and V1, Adoneta is the more highly specialized of the two genera.

It often happens that after the abortion of the base of media the distal cell is traversed by a more or less distinct line or scar indicating the former position of this part of the vein. This scar is indicated in the accompanying figure of the wings of Eacles (Fig. 18) by dotted lines.

It will be observed that the branches of media are not continuous with the branches of this scar. There has been a migration of the proximal end of the remaining portion of each branch towards the vein from which it gets its supply of air. Frequently there remain short stumps, projecting into the distal cell from the discal cross vein, and continuous with the scar, at the points where the branches formerly emerged from the discal cell. These are indicated by the arrows in Figure 13, and are also shown in Figure 18. It will be readily seen that the extent to which this migration of the base of a branch has gone will serve as an indication of the degree of divergence of the form from a primitive type.¹¹

¹¹ I wish here to acknowledge the assistance of Miss Clelia D. Mosher, who, while a student in my laboratory at Palo Alto, first worked out the relation of these stumps to the branches of media. 

In connection with this part of the discussion a few words regarding the nature of the so-called distal vein are appropriate. It is evident that this is not a single cross vein extending from radius to cubitus; but it is made up of several distinct elements, and these elements may differ in different genera. There is a cross vein between radius and the first branch of media, and another between the third branch of media and cubitus (Fig. 16, 18, c.v., c.v.). These extremes of the series forming the discal vein, however, have the appearance in many cases of being parts of longitudinal veins (Fig. 18); and in such cases have not been considered, heretofore, as parts of the distal vein. The intermediate portions of the discal vein may be merely the branches of media somewhat bent out of their primitive course. This condition is illustrated by the hind wings of Packardia (Fig. 16). Here the first branch of media has been drawn towards radius as a result of the change in the source of its air supply; and in a similar way the third branch of media, receiving a large part of its air from cubitus, is bent towards cubitus. In other cases cross veins have been developed between the branches of media, and these form part of the discal vein; this is the case where there is an interpolated cell in the distal cell (Fig. 12). Sometimes a part of the distal vein may be looked upon as a trail indicating the path along which the base of a branch of media has migrated. An instance of this kind can be seen in the fore wings of Eacles (Fig. 18). Here that part of the distal vein lying between the stump which is marked V1 and radius is the path over which the base of vein V1 has migrated.

The union of vein V1 with radius and of vein V3 with cubitus after the abortion of the base of media is what would be expected. But in which direction would one expect the base of vein V2 to migrate? Occupying an intermediate position between radius and cubitus it may go either way. It is like a stream in the middle of a level plain, a trifle may change its course. And thus we find that in some families it migrates towards cubitus, making this vein apparently four-branched, while in other families it goes towards radius, leaving cubitus apparently three-branched.

This difference may be looked upon as a difference in kind of specialization, and is frequently of high value as indicating a dichotomous division of the line of descent. It is obvious that in a family, where vein V2 has migrated far towards cu-bitus and has thus established its chief source of air supply in that direction, it is not probable that genera will arise in which vein V2 is more closely united to radius than to cubitus. To resume the figure, the plain through which the stream is flowing is an elevated plateau; a pebble may determine which of two slopes it shall descend; but when well started clown one, it cannot traverse the other.

This character, however, must be used with care. In families where the direction of the migration of the base of vein V2 has been firmly established, as in the Saturniidae (Fig. 18), and in the Lasiocampidae (Fig. 29), it is decisive. One need not hesitate a moment in determining to which of these two families a genus belongs. But there are other families in which the direction of this migration is not yet fixed; and here the character is of subordinate value.

Not only may the branches of one system of veins become joined to those of other systems as just described, but there are many forms in which two adjacent principal veins are coalesced to a greater or less extent. This occurs chiefly in the hind wings.

I will discuss the veins in regular order, beginning with costa. This vein is apparently wanting in the hind wings of most Lepidoptera, and but little can be said regarding the manner of its disappearance. It seems probable that in most cases it has simply become atrophied, the overlapping of the wings rendering it unnecessary or even undesirable. For when that stage in the development of the order was reached in which the two wings of each side overlapped to a considerable extent, was it not better that the costal margin of the hind wing should be flexible ? There was no longer any need of a stiff margin, this part of the wing being supported during the downward stroke by the overlapping part of the fore wing; while a flexible margin would act as a valve to prevent the escape of the air between the two wings. The two wings in this way present a continuous surface. In many moths there is a thickening of the basal part of the costal margin; this I believe to be the remnant of costa.

But although it seems probable that in many cases the costa of the hind wings has simply faded out leaving cell I to function as this costal valve, there are cases in which this valve is a precostal development, the costa having moved backwards and become consolidated with the subcosta. A good illustration of this is presented by the European genus Zygaena (Fig. 19). Here the costa and subcosta are distinct for a considerable distance, but become united into a single vein.

It will be observed that the basal portion of costa extends like a cross vein and forms a strong support for the frenulum. This part of costa is sometimes preserved when the remaining part is wanting. See figure of Castnia (Fig. 15).

In most genera of the Geometridae there is a faint indication of a remnant of costa extending from the humeral angle, at the base of the frenulum, to the subcosta, which is strongly angulated. The same thing is shown in Euphanessa, (Figs. 20, 21), which is probably a Geometrid genus.

In many of the Psychidae a remnant of costa is preserved (Figs. 22, 23). Here subcosta and radius are united for a considerable distance; then they separate and subcosta soon becomes joined to costa for a short distance. In Figure 23 I have represented what I believe to be the course of these three veins, slightly separating them where they are coalesced.

That part of subcosta that lies between its separation from radius and its union with costa appears like an oblique cross vein; and had the short spur that represents the terminal part of costa been lacking, its true nature would not have been suspected. Does not this arrangement of the veins in the Psychidae afford an explanation of the origin of the so-called intercostal vein which is characteristic of the Sphingidae? See Figure 24. In many butterflies the base of costa of the hind wings is preserved. This is well shown in Papilo (Pl. II, Fig. 2). This vein has been observed and figured by many writers; but it has always been considered a preeostal (i.e., humeral) vein. But I believe it is essentially different from the humeral vein or veins of the Lasiocampidae. The humeral veins of this family of moths are secondary developments; while the spur in the humeral angle of butterflies is a remnant of one of the primitive veins, the costa.

In the hind wings of many moths a coalescence of subcosta and radius also takes place to a greater or less extent. These two veins may be joined for a short distance, as in Packardia, (Fig. 16), or they may be merged into one for a considerable proportion of their length as in Megalopyge, (Fig. 25), and in the Sesiidae. Every degree of coalescence between these extremes exist.

It has been customary in cases like the last to consider the subcostal vein wanting; but it is only necessary to count the branches of the principal veins to see that all are present; moreover, in most cases the two coalesced veins are separate for a short distance near the base.

There are, however, forms in which the basal part of radius is wanting. This condition is brought about in this way. First, something interferes with the growth of the basal part of radius, and this vein becomes weaker than the other principal veins. This stage is exhibited by Prionoxystus, (Fig. 12), in which radius of the hind wings still persists but is much weaker than the other veins, except media. Correlated with this weakening of radius is the formation of a cross vein between it and subcosta (Fig. 12, c.v.). This is an establishment of a new source of air supply for the distal portion of radius, and renders less necessary the basal portion of that vein. Sometimes the two veins are drawn together, and the coalescence extends in both directions from the point of first union, resulting in the form presented by Megalopyge (Fig. 25); but in other cases that portion of radius between the point of union and the base of the wing becomes atrophied: An approach to this condition is shown by Acoloithis (Fig. 26).

In the fore wings where radius is branched, we often find an anastomosing of the branches. In this way are formed the accessory cells (Fig. 20). This anastomosing doubtless serves to strengthen the wing.

In the hind wings of all Lepidoptera, except Hepialis (Fig. 27) and Micropteryx (Fig. 28), all of the branches of radius are united into one. But the condition of radius in the two genera named shows that it is normally five-branched in the hind wings as well as in the fore wings.

In the discussion of media, given on a previous page, attention was called, so far as concerns its coalescence with other veins, merely to its branches; but the principal stem of this vein may become joined either to radius, as in the fore wing of Castnia (Fig. 15), or to cubitus, as in the fore wing of Prionoxystus (Fig. 12).

It will be observed that here is a character which is of value as indicating a dichotomous division of the line of descent.

I do not recall any instance where cubitus is coalesced with an anal vein to a marked degree, except in the Papilionidae (Pl. II, Fig. 2); but the growing together of different anal veins is a very common occurrence. This condition is preceded phylogenetically by the formation of a cross vein. Such a vein exists between veins IX and XI of the fore wings of Castnia (Fig. 15), and between veins VIII and IX of Thyridopteryx (Fig. 22).  Following this stage the two veins are drawn together; see veins IX and XI of the fore wings of Thyridopteryx (Fig. 22), and the same veins in Megalopyge (Fig. 25). Usually, however, when these veins are joined in this way, that part of vein XI beyond the point of union disappears, and vein IX presents the appearance of being forked towards the base. See Adoneta (Fig. 17).
In the fore wings of the Psychidae it frequently happens that the basal part of vein VIII disappears, and then vein IX appears to be forked outwardly (Fig. 22).

A good illustration of the coalescence of principal veins in another order of insects is presented by the dragon-flies (Odonata). Here veins III, IV and V are united into one from the base of the wing to the arculus. This coalescence is from the base of the wing outward, as is the coalescence of the main stem of media with either radius or cubitus in the Lepi-doptera. But most instances of coalescence in the Lepidoptera begin on the disc of the wing and extend in either direction. In the Diptera a third mode of coalescence is common. In this order it frequently happens that two longitudinal veins come together at their tips and unite, the coalescence proceeding from the margin of the wing towards the base. The result is that a cell which normally opens on the margin of the wing is closed at a greater or less distance before the margin; and the extent of this distance will be an indication of the degree of divergence from the primitive type.

The coalescence of two veins may be complete resulting in the reduction of the number of veins in the wing. This frequently happens especially with the branches of radius of the fore wings in the Lepidoptera. This vein is naturally five-branched; when a less number of branches occurs it is because the coalescence of some of the branches has proceeded to the margin of the wing.

The number of veins in the wing may be reduced, however, in another way: a vein may simply fade out. The most common instances of this kind in the Lepidoptera occur in the anal areas of the wings. As these areas become narrowed (i.e., specialized by reduction) one or two veins disappear.

The second branch of media is also a vein that is apt to disappear by atrophy; this occurs frequently in the Geometridae.

The usual result of specialization of the wings of Lepidoptera is a reduction of the number of veins where any change in the number is made. But it is not always so; for new veins may appear. I have already described the formation of cross veins, where a new source of air supply is established, and preceding the coalescence of distinct veins. In a somewhat similar way veins are formed in the basal part of cell I of the hind wings in the Lasiocampidae. In these cases the humeral angle has become greatly extended (Fig. 29). This outgrowth of the wing, like all other parts, is abundantly supplied wit h tracheae; and about some of the tracheae have been developed veins which protect them by stiffening this area so that it will not bend and thus compress them. This stiffening of the area doubtless serves another function to be described later.

These veins, developed in the humeral angle of the wing, I designate as the humeral veins. They have been termed the precostal veins; but the determination of the fact that the so-called costa of Lepidoptera is really the subcosta, renders the name precostal inappropriate.

The joining together of the two wings of each side in many moths by means of a frenulum and a frenulum hook, is a well known characteristic. But the real nature of the frenulum has not been understood, neither has its taxonomic value been appreciated.

I was led to make a careful study of this part of the wing by the discovery that in Hepialis an entirely different method of uniting the two wings of each side has been developed. In this genus, and as I have since discovered in Micropteryx also, instead of the wings being joined by a frenulum, which is a bristle or a bunch of bristles borne by the hind wing, they are joined by a membranous lobe extending back from near the base of the inner margin of the fore wing (Fig. 27, 28, j). To this lobe I have applied the name jugum.

When the wings of Hepialis are extended, the jugum projects back beneath the costal border of the hind wing, which, being overlapped by the more distal portion of the inner margin of the fore wing, is thus held between the two, as in a vice.

The discovery of the fact that there are two distinct modes of uniting the wings during flight suggests the inference that in the primitive Lepidoptera the wings were united in neither way. For it is not easy to see how one mode could have been developed from the other.

It is probable that in the primitive moths the mesothorax and metathorax were much more distinct than in the recent forms; and consequently the two pairs of wings were farther apart than now. As the consolidation of the thoracic segments advanced, the wings were brought nearer and nearer together, till finally the development of a connecting organ was rendered possible.

Such an organ might be borne by the fore wings, or it might be borne by the hind wings. In some moths the specialization took the former direction; in others, the latter; and thus arose a division of the order.

This division I consider of subordinal value; and I have already proposed the names Jugatae and Frenatae for the suborders thus indicated.*

* Proc. Am. Ass. Adv. Sci., Vol. XLI (1892), p. 200.

Let us try to obtain an idea of the ways in which the jugum and the frenulum were developed. As to the jugum I have but little to offer beyond the suggestion that at first it may have been merely an adventitious lobe, or a slight sinuosity in the inner margin of the fore wing. If such a lobe should project beneath the hind wing ever so little it would tend to insure the synchronous action of the two wings, and thus offer an opportunity for natural selection to act.

The frenulum is a much more complicated organ. As a rule we find that in the female it consists of several bristles, while in the male it consists of a single, strong spine. If one of the bristles of the compound frenulum of the female be examined it will be found to be hollow, containing a single cavity. But when the frenulum of a male is examined it is found to contain several parallel-cavities. Evidently the frenulum of the male. is composed of several bristles as is that of the female, but these bristles are grown together, forming a single strong spine. This can be easily seen by examining a bleached wing that has been mounted in balsam. Usually the cavities in the bristles contain air which renders them easily visible.

It is obvious, therefore, that the frenulum of the female exhibits a more generalized condition than does that of the male. In some females the frenulum is so slightly developed that the bristles composing it are little more than hairs. This fact suggests that the primitive frenulum was developed from a bunch of hairs, clothing the base of the wing. Such a tuft of hairs projecting under the fore wing would tend in a slight degree to insure the synchronous action of the two wings; and as soon as these hairs had assumed this function the tendency of natural selection would be to strengthen them. In the female of Prionoxystus the frenulum consists of a series of bristles which vary in size from a short hair to a comparatively long spine (Fig. 12, f). This throws much light on the development of this organ.

The extent to which the specialization of the frenulum has been carried is remarkable. In the males of some of the Psychidae it is a strong spine nearly half as long as the hind wing (Fig. 22). In the Cymatophoridae it is furnished with a knob at the tip. But the most remarkable feature of this specialization is the development of a membranous fold on the fore wings of males for receiving the end of the frenulum, and thus more securely fastening the two wings together.

This fold, or frenulum hook, is so well known that it is unnecessary to enter upon a detailed discussion of it. I will, therefore, merely record a few observations that I have made upon it. In all families in which I have observed it, it arises from the membrane of the wing near the base of cell I (Fig. 22), except that in Castnia it seems to have been pulled back so that it arises from the subcostal vein. The uniformity in the position of the frenulum hook indicates that it was developed before those families in which it exists had become separated phylogenetic-ally. For if it had been independently developed in the different families there would probably have been a lack of uniformity in its position.

Some light is thrown upon the probable origin of the frenulum hook by the fact that in many females there is a tuft of curved scales projecting back from the base of cell I, and serving to hold the frenulum in place. In many moths there is also a tuft of scales projecting forward from the base of cell VII, which functions in a similar way.

In certain families of moths (Saturniina, Lasiocampidae, and Drepanidae) and in all butterflies there is neither a frenulum nor a jugum. But in other respects the wing characters of these moths and of butterflies agree quite closely with those of the frenulum-bearing moths, and do not agree with the Jugatae in their distinctive characters (i.e., in an equal reduction of the two pairs of wings, and in having radius of the hind wings branched).

If the wings of one of these moths or of a butterfly be examined it will be seen that there is a large expansion of the humeral angle of the hind wings (Figs. 30, 14), which causes the two wings of each side to overlap to a much greater extent than they do in other Lepidoptera.

This extensive overlapping of the wings effectually insures their synchronous action without the aid of a frenulum, and I believe explains the loss of the frenulum. This theory is supported by the fact that in the more generalized genera of the Saturniina (Perophora and Sericaria) where the humeral angle is not expanded to so great a degree as it is in the more specialized forms, there remains a rudiment of the frenulum, (Figs. 31, 32). And in the Drepanidae where the frenu-lum is usually wanting, it persists in one sex in certain genera.

It is important that this expanded humeral angle should have a certain degree of stiffness if it is to perform the function of a frenulum. This has been obtained in some cases by a more or less diffused thickening of the membrane of the wing. Such a thickening is represented by the dotted portion in the figure of the hind wing of Eacles, (Fig. 18). In other cases the thickening takes place along a definite line and encloses a trachea; thus are formed the humeral veins of the Lasiocampidae, (Fig. 29).

An interesting fact in connection with this abortion of the frenulum, is that in Perophora the rudiment of the frenulum of the male consists of a bunch of bristles. This is an excellent illustration of an organ which, in the course of its abortion, retraces the steps by which it was formed. In Sericaria the bristles composing the rudimentary frenulum in the male are still consolidated.

This modification of certain hairs on the costa of the hind wing into an organ whose function is to fasten the two wings together, is paralleled by the development of a row of hooks on the costa of the hind wings in the Hymenoptera and in the Aphididae, which has a similar function. And the development of a jugum has taken place in the Trichoptera. In fact in several respects the Trichoptera and the Jugatae resemble each other more closely than do the Jugatae and the Frenatae.

When a careful study is made of the wings of the two sexes of a species it often happens that a marked difference is found in them; and so far as I have observed the difference indicates a higher degree of specialization on the part of the male. It seems as if the female lagged behind the male in the race for perfection of organs. This is often shown in the degree to which the branches of the veins are consoli dated. But it is shown most markedly in the structure of the frenulum as already pointed out.
 
The explanation of this comparative lack of specialization of the wings in females is to be found largely, I believe, in the fact that the males seek their mates, while the females await the approach of the males. Many instances are well known (Orgyia, Anisopteryx, et al.) where the females have lost their wings through disuse while the males retain well developed wings. The only instance that I can call to mind where the reverse has occurred, is the case of Blastophaga. Here the male has no need of wings, as he finds his mate in the cavity of the fig in which he has been developed; while the female must fly elsewhere to deposit her eggs in a suitable place.

The great difference in the habits of flight of the two sexes in many moths is well illustrated by the results of a series of experiments with trap lanterns which I conducted several years ago. Six lanterns were kept burning from spring to fall, and each day's catch was kept separate. The results have been partially tabulated by Mr. Slingerland, and I extract the following table from his report.

Feltia jaculifera is not a very common species at Ithaca and hence the results obtained with this species are not so important as those obtained with Feltia subgothica. This is our most common Noctuid; and of the specimens captured (2,382 in all) more than 94 per cent. were males. It is not at all likely that this represents the difference in the numbers of individuals of the two sexes; it is much more probable that the difference is due to a greater activity on the part of the males.

While I believe that the greater specialization of the wings of the male is due to the greater activity of that sex, I confess that I am greatly puzzled by the fact that in no female of the Macrofrenatae has a consolidated frenulum and a frenulum hook been developed.* In other respects the females, as a rule, lag behind the males in their specialization only a short distance. But while the possession of a consolidated frenulum and a frenulum hook is attained by the males in the most generalized of living frenate moths, that stage is not reached by  the females of the most specialized genera. Obviously there is a factor here that I have not discovered.

In this discussion of the taxonomic value of the wings, I have confined myself chiefly to a study of the form of the wings, their venation, and the relation of the two pairs to each other. But I believe that even the clothing of the wings is of great taxonomic value.
 

If the scales of any of the more generalized moths be examined, they will be found to be long and narrow and scattered irregularly over the surface of the wing. On the other hand in the more specialized members of the order, as in most butterflies, the scales are much less hair-like, being short and broad; and they are arranged in regular overlapping rows. Evidently both the form of the scales and their arrangement upon the wing offer indications as to the degree of divergence from a primitive type of the insect bearing them.

More than this I am convinced that in some eases at least the form of the scales is characteristic of a particular line of development. One can determine, for example, without any doubt whether a moth belongs to the Lasiocampidae or not by merely examining the scales of the wings.

I was impressed with the taxonomic value of the scales very soon after I began the systematic study of the Lepidoptera according to the method outlined in this essay. But the time at my disposal would not admit of my investigating this part of the problem in a satisfactory manner; and at my request the investigation has been undertaken by my colleague Professor V. L. Kellogg of the Leland Stanford Junior University.

Professor Kellogg is preparing an elaborate paper on this subject, which will be published soon after the appearance of the volume containing this one.

While the chief object which I have had before me is the indication of a method of taxonomic work, I hope this essay will be of value to entomologists in hastening the adoption of a uniform nomenclature of the parts of the wings of insects, and thus make easier the study of the relation of the different orders of insects to each other. In order that this nomenclature may be more complete I propose the following method of naming the cells of the wing; for we have as yet no system that is of general application.

The method I propose is, briefly, to designate each cell by the name or number of the vein that forms its front margin when the wings are spread.

The application of this system to the Lepidoptera is indicated by Figure 33.
 
In certain special cases special names may be desirable.  Thus in the Lepidoptera the cells formed by the anastomosing of the branches of radius are commonly known as the accessory cells, which is a very convenient term. And cell III + V is universally known as the discal cell. In some of the more generalized moths a cell is formed within the distal cell by a forking of media: this has been termed the interpolated cell (Figs. 12, 22, 27).

In those orders where there are transverse veins, each of the cells between two longitudinal veins is divided into a series of cells, and can be so designated. Thus the series of cells between veins III1 and III2 may be termed the first series of radial cells; and the members of such a series can be numbered. If one should speak of the peculiar form of the third cell in the second radial series, there need be no difficulty in determining the cell indicated, even by one who had not made a special study of the order to which the insect in question belongs. Heretofore it has been necessary for the student to learn a distinct nomenclature for each order, and in some cases for each family, studied.

In concluding this part of this essay I wish to refer to two curious methods of specialization that have interested me greatly. In certain cases where the body of the insect has become greatly reduced in size, a reduction of the area of the wing membrane has taken place and correlated with this there has been a great expansion of the fringe of the wing. The best known examples of this are the narrow-winged Tineids, the Thysanoptera, and certain parasitic Hymenoptera. This kind of specialization seems possible only with minute insects, where the weight to be supported during flight is not great.

In a Tineid which I have studied the hairs composing the fringes of the wing are inserted in the lower side of the wing-membrane a short distance back from the edge of the wing; and the edge of the wing is stiffened above by strong overlapping scales. This arrangement renders the fringes rigid during the downward stroke of the wing, but admits of their depression during the upward stroke; a combination well adapted to facilitate flight.

The second method of specialization referred to above is the loss of the front wings in the Coleoptera and Euplexoptera. In these two orders the paraptera of the mesothorax have been developed into elytra, and have crowded out the front pair of wings. The function of flight has been relegated in this way to the hind wings.

This homology of the elytra of beetles with the tegulae of Hymenoptera and with the patagia of Lepidoptera was pointed out by F. Meinert long ago*; but Meinert's paper seems to have escaped the attention of entomologists almost entirely. It is referred to by C. Hoffbauer in his paper on the minute structure of the elytra. ^$ But although Hoffbauer shows conclusively that the structure of the elytra resembles that of the pronotum and differs in every essential feature from that of the wings, strangely enough he does not accept the conclusion of Meinert.

Meinert also pointed out the fact that in many Coleoptera (e.g., Dytiscus) rudiments of the front wings exist beneath the elytra.

* Meinert, F. (1880) Sur l'Homologie des Elytres des Coleopteres. Entomologisk Tidskrift 1:168.

$ Hoffbauer, G. (1892) Beitrag zur Kenntnis der Insektenfluegel. Zeit.Wiss.Zool. 54:579-630.

As to the cause of this strange specialization I can only conjecture that in the primitive Coleoptera the habits of the insects were such that the protection of the wings by elytra was of more importance than that the first pair should be functional; cephalization was sacrificed in order that the remaining pair of wings might be protected. It may be that the primitive Coleoptera were wood borers, the only paleozoic remains supposed to be of beetles are borings; or their habits may have been like those of the recent Carabidae. In either case the wings would be in need of special protection.